Match can return non-boolean results yet warning suggests results are boolean

[fuglede]

Code Sample, a copy-pastable example if possible

In [26]: pd.Series(['a', 'b', 4]).str.match('a')
Out[26]:
0     True
1    False
2      NaN
dtype: object

In [27]: pd.Series(['a', 'b', 4]).str.match('a', as_indexer=True)
# C:\Users\username\AppData\Local\Continuum\Anaconda3\Scripts\ipython:1: FutureWarning: 'as_indexer' keyword was specified but is ignored (match now returns a boolean indexer by default), and will be removed in a future version.
Out[27]:
0     True
1    False
2      NaN
dtype: object

Problem description

The warning states that match (and similarly for search) returns a boolean indexer by default, yet that's not the case for numeric input, as the example shows.

This one caught me off guard as it caused an indexing error in a piece of code that assumed that the result of match could be used as a mask regardless of input types.

Expected Output

The output currently obtainable by letting na=False:

In [33]: pd.Series(['a', 'b', 4]).str.contains('a', na=False)
Out[33]:
0     True
1    False
2    False
dtype: bool

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.5.final.0
python-bits: 64
OS: Windows
OS-release: 8.1
machine: AMD64
processor: Intel64 Family 6 Model 61 Stepping 4, Genuin
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None

pandas: 0.23.4
pytest: 3.3.2
pip: 18.0
setuptools: 38.4.0
Cython: 0.27.3
numpy: 1.13.3
scipy: 1.0.0
pyarrow: None
xarray: None
IPython: 6.2.1
sphinx: 1.6.3
patsy: 0.5.0
dateutil: 2.6.1
pytz: 2017.3
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.4
feather: None
matplotlib: 2.1.2
openpyxl: 2.4.9
xlrd: 1.1.0
xlwt: 1.3.0
xlsxwriter: 1.0.2
lxml: 4.1.1
bs4: 4.6.0
html5lib: 0.9999999
sqlalchemy: 1.2.8
pymysql: None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[TomAugspurger]

That keyword is slated to be removed in 0.24 #6581

Could you make a PR just removing it, rather than changing the error message?

[HyunTruth]

@TomAugspurger I created a PR with the removal and change of test to accomodate the change #22356

[fuglede]

Heh, one has to be fast around here.

[jreback]

maybe my search is not working properly (as this is not listed as all), but this is versionchanged in 0.21.0 on the change, hence we would remove > 0.24, @TomAugspurger ?

[TomAugspurger]

#6581 has it listed as under (0.24 / 1.0). Not sure what that means.